Text Generation
GGUF
edge0
Mixture of Experts
edge-inference
latent-moe
4-bit precision
agent
coding
cuda
conversational
Instructions to use Edge0/Edge0-145B-q3k with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use Edge0/Edge0-145B-q3k with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf Edge0/Edge0-145B-q3k:BF16 # Run inference directly in the terminal: llama cli -hf Edge0/Edge0-145B-q3k:BF16
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf Edge0/Edge0-145B-q3k:BF16 # Run inference directly in the terminal: llama cli -hf Edge0/Edge0-145B-q3k:BF16
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf Edge0/Edge0-145B-q3k:BF16 # Run inference directly in the terminal: ./build/bin/llama cli -hf Edge0/Edge0-145B-q3k:BF16
Use Docker
docker model run hf.co/Edge0/Edge0-145B-q3k:BF16
- Llama new
- LM Studio
- Jan
- vLLM
How to use Edge0/Edge0-145B-q3k with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Edge0/Edge0-145B-q3k" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Edge0/Edge0-145B-q3k", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Edge0/Edge0-145B-q3k:BF16
- Ollama
How to use Edge0/Edge0-145B-q3k with Ollama:
ollama run hf.co/Edge0/Edge0-145B-q3k:BF16
- Unsloth Desktop
- Pi
How to use Edge0/Edge0-145B-q3k with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Edge0/Edge0-145B-q3k:BF16
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:9931/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "Edge0/Edge0-145B-q3k:BF16" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use Edge0/Edge0-145B-q3k with Docker Model Runner:
docker model run hf.co/Edge0/Edge0-145B-q3k:BF16
- Lemonade
How to use Edge0/Edge0-145B-q3k with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Edge0/Edge0-145B-q3k:BF16
Run and chat with the model
lemonade run user.Edge0-145B-q3k-BF16
List all available models
lemonade list
- Hermes Agent
How to use Edge0/Edge0-145B-q3k with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Edge0/Edge0-145B-q3k:BF16
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:9931/v1 hermes config set model.default Edge0/Edge0-145B-q3k:BF16
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use Edge0/Edge0-145B-q3k with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Edge0/Edge0-145B-q3k:BF16
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:9931/v1 \ --custom-model-id "Edge0/Edge0-145B-q3k:BF16" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Commit History
Yesterday
-
docs: say NVFP4 runs on Metal too, and list the macOS build
Verified signature4bebcc0 -
docs: say where strata-vision comes from
Verified signature854d962 -
docs: say where strata-vision comes from
Verified signaturead810e0 -
docs: add 12 GiB VRAM + 20 GiB RAM speed row and small-card setup; merge low-RAM blocks
Verified signatureb10b58e
5 days ago
-
docs: say where strata-vision comes from (not the Edge0 source tree)
Verified signaturea19f88c -
docs: say what the vision encoder is and how to turn it on
Verified signaturebd46f1d -
Drop the old encoder filename
Verified signature200eac5 -
Add the vision encoder as mmproj-BF16.gguf
Verified signature10fc212 -
Add the vision encoder (mmproj-Qwen3.8-Flash-Next-BF16)
Verified signature839a139 -
probe: cleanup
Verified signature266669d -
probe: xet speed
Verified signature34340c7 -
probe: cleanup
Verified signaturebec855e -
probe: mac path check
Verified signaturefd2c858 -
probe: cleanup
Verified signaturef30dfb5 -
probe: upload path check
Verified signature4bba072 -
docs: spell out the Q3_K record (Q3_K gate/up + IQ4_NL down)
Verified signature4136439 -
docs: compare the expert pool at FP4 on both sides
Verified signature5a2a140 -
docs: add rebuilt-not-converted section
Verified signaturee43f976
6 days ago
-
docs: macOS (Apple Silicon) support
Verified signature2d80116
7 days ago
-
docs: drop the reserve arithmetic from the measurement conditions
Verified signature00f91ca -
docs: drop the editorial clause about cgroups
Verified signatureebcbad3 -
docs: how the 16 GiB VRAM budget is enforced; add --spec to the VRAM table
Verified signature1b1dbf7 -
docs: per-pack env in the capacity-controls example (q3k vs NVFP4)
Verified signature6c25b17 -
docs: blank line so the capacity-controls heading renders
Verified signatured3f6611 -
docs: drop the unmeasured 16 GiB worked example from capacity controls
Verified signature6e6d1cd -
docs: drop residency-conditions sentence, add capacity controls section
Verified signature117c71d -
Add low-memory startup argument example
Verified signatureReplace Combined mode with concrete expert-cache arguments and a static profile example. Preserve all other content, assets and repository settings.82e960c -
Add compact memory usage help and measurement conditions
Verified signatureAppend usage help after Quick start and clarify residency measurement settings. Preserve every existing README byte, asset and repository setting.dd6864e
8 days ago
-
Clarify model licensing and retain Qwen upstream terms
Verified signatureDesignate Edge0-specific rights under the Edge0 Community Model License 1.0; retain the Qwen license verbatim. Change only model-card license fields and Credits. Preserve all other page content, styling, assets, model files, disclosure states and repository settings. Existing valid grants remain preserved.d7de835 -
Expand all model card sections by default
Verified signatureOnly add open attributes to existing details elements. Preserve all current online text, styling, links, data, assets and repository settings.88865e1 -
Update README.md
Verified signature0e932fc -
Publish the reviewed typography, colored speed table and independent video section
Verified signatureInclude the same scoped card styles and complete rendered content used locally. Preserve all prose, scores, measurements, links and disclosure defaults. No model weights, configuration, visibility or video changes.1823df2 -
Preserve two-column introduction and inline badges on Hugging Face
Verified signatureKeep the approved wording in a standalone SVG, use HF-recognized badge filenames, and fix full-size figure URLs. No weight, configuration, visibility or video changes.0a6b9b2 -
Add Edge0 model card and presentation assets
Verified signaturePublish the prepared model card, diagrams, badges and demo references. The shared demo MP4 is hosted in Edge0-145B-NVFP4.3f01735
9 days ago
-
Update model.json model_name to edge0-145b-q3k
Verified signature07b1c3b -
Upload Edge0-145B-q3k
Verified signature3a62b19 -
initial commit
Verified signature1b80fd8